CR Comments (3e40b41c) · Commits · 钟慕尧 / deepchem

contrib/atomicconv/feat/atomicnet_coordinates.py

+8 −8

Original line number	Diff line number	Diff line
		@@ -13,7 +13,7 @@ import numpy as np
		from rdkit import Chem
		from deepchem.feat import Featurizer
		from deepchem.feat import ComplexFeaturizer
		from deepchem.utils import pad_array
		from deepchem.utils import pad_array, rdkit_util


		def get_cells(coords, neighbor_cutoff):
		@@ -153,9 +153,8 @@ def compute_neighbor_cell_map(N_x, N_y, N_z):
		for x_offset in offsets:
		for y_offset in offsets:
		for z_offset in offsets:
		neighbors.append(
		((x_ind + x_offset) % N_x, (y_ind + y_offset) % N_y,
		(z_ind + z_offset) % N_z))
		neighbors.append(((x_ind + x_offset) % N_x, (y_ind + y_offset) %
		N_y, (z_ind + z_offset) % N_z))
		neighbor_cell_map[(x_ind, y_ind, z_ind)] = neighbors
		return neighbor_cell_map

		@@ -245,6 +244,7 @@ class NeighborListAtomicCoordinates(Featurizer):
		Molecule

		"""
		print(mol)
		N = mol.GetNumAtoms()
		coords = get_coords(mol)

		@@ -394,10 +394,10 @@ class ComplexNeighborListFragmentAtomicCoordinates(ComplexFeaturizer):
		"""

		try:
		frag1_mol = Chem.MolFromPDBFile(
		frag1_pdb_file, sanitize=False, removeHs=False)
		frag2_mol = Chem.MolFromPDBFile(
		frag2_pdb_file, sanitize=False, removeHs=False)
		frag1_mol = rdkit_util.load_molecule(
		frag1_pdb_file, add_hydrogens=False, calc_charges=False)[1]
		frag2_mol = rdkit_util.load_molecule(
		frag2_pdb_file, add_hydrogens=False, calc_charges=False)[1]
		except:
		frag1_mol = None
		frag2_mol = None

+2 −1

Original line number	Diff line number	Diff line
		@@ -9,6 +9,7 @@ from __future__ import unicode_literals
		import os
		import numpy as np
		import pandas as pd
		import deepchem as dc
		from atomicnet_coordinates import ComplexNeighborListFragmentAtomicCoordinates


		@@ -68,7 +69,7 @@ def compute_pdbbind_coordinate_features(complex_featurizer, pdb_subdir,
		"""

		protein_file = os.path.join(pdb_subdir, "%s_pocket.pdb" % pdb_code)
		ligand_file = os.path.join(pdb_subdir, "%s_ligand.pdb" % pdb_code)
		ligand_file = os.path.join(pdb_subdir, "%s_ligand.sdf" % pdb_code)
		feature = complex_featurizer._featurize_complex(
		str(ligand_file), str(protein_file))
		return feature

+5 −8

Original line number	Diff line number	Diff line
		@@ -9,16 +9,13 @@ __license__ = "MIT"
		import os
		import sys
		from subprocess import call
		from atomicnet_pdbbind_datasets import load_core_pdbbind_fragment_coordinates
		from atomicnet_pdbbind_datasets import load_pdbbind_fragment_coordinates

		call([
		"wget",
		"wget", "-c",
		"http://deepchem.io.s3-website-us-west-1.amazonaws.com/datasets/pdbbind_v2015.tar.gz"
		])
		call(["tar", "-xvzf", "pdbbind_v2015.tar.gz"])

		# This could be done with openbabel in python
		call(["convert_ligand_sdf_to_pdb.sh"])
		# call(["tar", "-xvzf", "pdbbind_v2015.tar.gz"])

		base_dir = os.getcwd()
		pdbbind_dir = os.path.join(base_dir, "v2015")
		@@ -30,6 +27,6 @@ complex_num_atoms = 908
		max_num_neighbors = 8
		neighbor_cutoff = 12.0

		pdbbind_tasks, dataset, transformers = load_core_pdbbind_fragment_coordinates(
		pdbbind_tasks, dataset, transformers = load_pdbbind_fragment_coordinates(
		frag1_num_atoms, frag2_num_atoms, complex_num_atoms, max_num_neighbors,
		neighbor_cutoff, pdbbind_dir, base_dir, datafile)
		neighbor_cutoff, pdbbind_dir, base_dir, str(datafile))

+1 −1

Original line number	Diff line number	Diff line
		@@ -242,7 +242,7 @@ class Weights(Input):
		super(Weights, self).__init__(**kwargs)


		class LossLayer(Layer):
		class L2LossLayer(Layer):

		def __init__(self, **kwargs):
		super(LossLayer, self).__init__(**kwargs)

+13 −21

Original line number	Diff line number	Diff line
		@@ -231,17 +231,17 @@ class TensorGraph(Model):
		retval = np.expand_dims(from_one_hot(retval, axis=2), axis=1)
		return retval

		def predict_proba_on_generator(self, generator, sess=None):
		def predict_proba_on_generator(self, generator):
		"""
		TODO: Do transformers even make sense here?

		Returns:
		y_pred: numpy ndarray of shape (n_samples, n_classes*n_tasks)
		"""
		if not self.built:
		self.build()

		def predict_closure(session):
		with self._get_tf("Graph").as_default():
		with tf.Session() as sess:
		saver = tf.train.Saver()
		saver.restore(sess, self.last_checkpoint)
		out_tensors = [x.out_tensor for x in self.outputs]
		results = []
		for feed_dict in generator:
		@@ -249,20 +249,12 @@ class TensorGraph(Model):
		self.layers[k.name].out_tensor: v
		for k, v in six.iteritems(feed_dict)
		}
		result = np.array(session.run(out_tensors, feed_dict=feed_dict))
		result = np.array(sess.run(out_tensors, feed_dict=feed_dict))
		if len(result.shape) == 3:
		result = np.transpose(result, axes=[1, 0, 2])
		results.append(result)
		return np.concatenate(results, axis=0)

		if sess is not None:
		return predict_closure(sess)
		with self._get_tf("Graph").as_default():
		with tf.Session() as sess:
		saver = tf.train.Saver()
		saver.restore(sess, self.last_checkpoint)
		return predict_closure(sess)

		def predict_on_batch(self, X, sess=None):
		"""Generates output predictions for the input samples,
		processing the samples in a batched way.